An argument injection vulnerability exists in the ad-hoc RRD graph feature of multiple versions of OpenNMS Meridian and Horizon. The ad-hoc graph endpoint (GET /graph/graph.png?adhoc=true) copies request parameters into the rrdtool command line without validation. A low-privileged authenticated user (ROLE_USER or ROLE_DASHBOARD) can insert extra rrdtool arguments. An injected --daemon argument makes the server connect to an attacker-chosen address, which is a server-side request forgery. An injected DEF statement reads a local RRD file by absolute path. The command runs through exec with an argument array and no shell, so arbitrary operating system command execution is not possible. The fix validates the ad-hoc graph parameters against allowlists.
The solution is to upgrade to Meridian 2024.3.15, 2025.0.12, 2026.0.2 and Horizon 36.0.6 or newer. Meridian and Horizon installation instructions state that they are intended for installation within an organization's private networks and should not be directly accessible from the Internet.
A vulnerability in obsidian-tc versions 1.13.0 through 1.31.3 allows improper authorization of certain destructive vault operations. When deployed with Streamable HTTP and auth.jwtSecret configured, a client supporting elicitation.form could bypass human approval by resubmitting a valid requestState, causing the server to trust the approval state incorrectly. This affects only tools marked as destructive or requiring human-in-the-loop scope. The issue is fixed in version 1.31.4.
CVE-2026-55606 is a buffer over-read vulnerability in the odhcpd DHCP server component used by OpenWrt. In versions 25.12 and earlier, a flaw in DHCPv6 self-loop detection causes memcmp() to read beyond the intended buffer size on little-endian systems. This can be triggered by a single unauthenticated DHCPv6 packet, leading to a 2560-byte overread from a 128-byte stack buffer. No patches are publicly available at the time of disclosure.
Intruders used compromised credentials to access an Educational Testing Service (ETS) SFTP server, copying records of California public school students dating back to 2006. The breach was discovered on September 25, 2026, but unauthorized access occurred intermittently from May 22 to September 18, 2026. The exposed data includes personal information such as student names, birth dates, addresses, school details, and sensitive educational data, but excludes Social Security numbers, passwords, financial information, and parent/guardian data. ETS disabled the compromised account and took the SFTP site offline, working on a replacement with stronger security. Notifications and identity protection services are being provided to affected individuals. The exact method of credential compromise and attacker identity remain unknown.
obsidian-tc is a governed, agent-ready Model Context Protocol server for Obsidian vaults. Prior to 1.3.1, bridge tools do not consistently apply the read access-control list (ACL) to every returned field. The tasks_filter tool exposes group counts derived from denied notes, makemd_query passes unfiltered sibling result fields through, and list_templates exposes template paths and parsed user-function bodies without a read-whitelist check. An authenticated caller can consequently infer or read data from notes outside the caller's read ACL. Version 1.3.1 fixes the vulnerability by filtering path-attributable fields and applying the read whitelist to list_templates.
A JEXL expression sandbox bypass exists in multiple versions of OpenNMS Meridian and Horizon. A low-privileged authenticated user can submit a crafted expression to the Measurements REST API that escapes the sandbox and loads arbitrary Java classes on the server. This can potentially allow an attacker to gain access to confidential information and compromise integrity.
The solution is to upgrade to Meridian 2024.3.12, 2025.0.9 and Horizon 36.0.3 or newer. Meridian and Horizon installation instructions state that they are intended for installation within an organization's private networks and should not be directly accessible from the Internet.
The npm package 'approval-guardian' masquerades as an on-chain approval scanner but contains malicious code that silently harvests credentials. When invoked via CLI, it scans user directories for wallet application data and browser extension settings related to cryptocurrency wallets and password managers. It also includes a Windows keylogger implemented via PowerShell/C# to capture sensitive input such as MetaMask unlock passwords. The stolen data is sent to a hardcoded command-and-control server. Multiple specific versions of the package are affected.
For the latest discoveries in cyber research for the week of 27th July, please download our Threat Intelligence Bulletin. TOP ATTACKS AND BREACHES Minnesota IT Services has confirmed coordinated cyberattacks affecting more than 30 community water utilities across the state. The incidents briefly disrupted a treatment plant in Braham and affected industrial control systems. Officials reported that drinking water safety was not affected. While the attack was not officially attributed, federal officials previously posted warning regarding targeting of critical infrastructure by Iranian-affiliated threat actors. Bank of Baroda, a major Indian bank, has disclosed an email account compromise that exposed internal communications and attachments. Reports claim more than 700GB of customer files, loan documents, and audit records were leaked, although the bank has not confirmed the reported volume. Core banking systems were unaffected. Amgen, a US biotechnology company that develops medicines for serious illnesses, has confirmed a breach involving cloud environments operated by third-party providers. Attackers exfiltrated proprietary corporate information and patient health data. The company reported no disruption to manufacturing, financial reporting, products, or its ability to supply medicines. Angola’s largest telecommunications provider, Unitel, has suffered a cyberattack that disrupted voice, mobile data, and internet services for millions of customers. The outage also affected electronic payments shortly before the company’s stock market debut. Network data indicated that internal systems were disabled while external routers remained online. AI THREATS Anthropic has disclosed that Claude-based cybersecurity models gained unauthorized access to systems belonging to three outside organizations during controlled evaluations. The models moved beyond intended test environments and reached sensitive production assets. Anthropic identified the incidents while reviewing testing practices following separate autonomous AI security failures. Researchers have published details of CVE-2026-59726, a critical vulnerability in the Ruflo AI agent platform. An unauthenticated attacker could abuse its exposed Model Context Protocol bridge to execute commands, steal API keys, access conversations, and alter stored AI memory. Ruflo addressed the issue in version 3.16.3. Researchers surfaced a privacy issue in Anthropic’s Claude sharing feature that allowed publicly shared conversations and artifacts to be indexed by search engines. Indexed content reportedly included personal information, resumes, financial records, access codes, API keys, and clinical trial material that users may not have expected to become searchable. VULNERABILITIES AND PATCHES Cisco has addressed CVE-2026-20316, an actively exploited vulnerability in Secure Firewall Management Center. The flaw allows unauthenticated attackers to access a built-in low-privileged account and retrieve sensitive information from affected systems. Cisco released hotfixes after exploitation was identified, and the vulnerability was added to CISA’s catalog. Broadcom has released patches for five vulnerabilities affecting VMware vCenter, ESX, Workstation, and Fusion. Three critical flaws could allow authentication bypass, arbitrary code execution, or escape from a virtual machine to its host. The issues include CVE-2026-59309 and CVE-2026-59310, both carrying CVSS scores of 9.8. JetBrains has released fixes for CVE-2026-63077, a critical authentication bypass affecting all TeamCity On-Premises versions. A remote unauthenticated attacker could execute code with TeamCity server privileges and compromise connected build environments. The flaw is fixed in versions 2025.11.7 and 2026.1.3. TeamCity Cloud was not affected. Rails maintainers have patched CVE-2026-66066, a critical Active Storage vulnerability affecting applications that use libvips. An unauthenticated attacker could read sensitive server files…
When the upstream Guardian or CMC was configured in the Remote Collector via n2os-tui, the generated configuration disabled TLS certificate verification, and no option was provided to enable it. A malicious actor could perform a man-in-the-middle attack and intercept the communication between the Remote Collector and the Guardian or CMC. This could result in theft of the sync token, impersonation of the server, injection of spoofed data (such as false asset information or vulnerabilities) into the Guardian or CMC, or disruption of the data flow between the Remote Collector and the Guardian or CMC.
Guardian language-system passes the id GET parameter directly into a PHP exec() call in text_to_subtitles.php (line 19) without sanitization: exec(\"php jobs/text_to_subtitles.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to execute arbitrary OS commands on the server.
Guardian language-system passes the id GET parameter directly into a PHP exec() call in transcribe.php (line 15) without sanitization: exec(\"php jobs/transcribe.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to execute arbitrary OS commands on the server.
Guardian language-system passes the id GET parameter directly into a PHP exec() call in transcribe_amazon.php (line 15) without sanitization: exec(\"php jobs/transcribe_amazon.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to execute arbitrary OS commands on the server.
Guardian language-system passes the id GET parameter directly into a PHP exec() call in translate_text.php (line 18) without sanitization: exec(\"php jobs/translate_text.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to execute arbitrary OS commands on the server.
Guardian language-system passes the id GET parameter directly into a PHP exec() call in speech_text.php (line 18) without sanitization: exec(\"php jobs/speech_audio_text.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to execute arbitrary OS commands on the server.
Guardian language-system passes the id GET parameter directly into a PHP exec() call in speechmac.php (line 18) without sanitization: exec(\"php jobs/speech_audio_mac.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to execute arbitrary OS commands on the server.
Guardian language-system passes the id GET parameter directly into a PHP exec() call in speechmac_text.php (line 18) without sanitization: exec(\"php jobs/speech_audio_mac_text.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to execute arbitrary OS commands on the server.
Guardian language-system passes the id GET parameter directly into a PHP exec() call in complex_start.php (line 14) without sanitization: exec(\"php jobs/complex.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to execute arbitrary OS commands on the server.
Guardian language-system passes the id GET parameter directly into a PHP exec() call in speech.php (line 18) without sanitization: exec(\"php jobs/speech_audio.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to execute arbitrary OS commands on the server.
Guardian language-system passes the id GET parameter directly into a PHP exec() call in text.php (line 15) without sanitization: exec(\"php jobs/text.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to execute arbitrary OS commands on the server.
Guardian language-system passes the id GET parameter directly into a PHP exec() call in translate.php (line 14) without sanitization: exec(\"php jobs/translate.php \".$login_session.\" \".$_GET['id'].\" ...\"). No authentication is required. An unauthenticated remote attacker can append shell metacharacters to execute arbitrary OS commands on the server.